Empowering Mental Health: CNN and LSTM Fusion for Timely Depression Detection in Women

Abstract

Depression is a mental illness that manifests as persistent melancholy, a loss of interest in routine activities, trouble focusing, poor memory, and a lack of energy. It is a widespread mental health condition that can affect people of any age or gender. Depression is more common in women than in men. In order to identify early indicators of depression in women, this study uses a deep learning-based model utilizing convolutional neural networks and Long Short-Term Memory. With the help of a dataset of left and right hemispheric electroencephalogram data, the suggested model was trained and assessed. The suggested method entails preprocessing the electroencephalogram data, which is feature-extracted using a convolutional neural network, and sequence modeling using a Long Short-Term Memory network. With the help of Electroencephalogram data from women with and without depression, the model was trained and assessed. The results show that the suggested method successfully identified depression in women using Electroencephalogram data with excellent accuracy, sensitivity, and specificity. When it came to identifying female depression, the model had an accuracy of 99.02% on the left hemisphere and a right hemisphere accuracy of 98.06%. The study demonstrates that employing advanced deep learning techniques on electroencephalogram data enables accurate and sensitive identification of depression in women. This highlights the potential for early intervention in mental health disorders, particularly in populations with a higher depression prevalence.